Re: [PATCH] SourceLevelDebugPkg/DebugAgentCommon: Use global variable for revision

["Kinney, Michael D" <michael.d.kinney@intel.com>]

Ard,

I agree.  Adding const would be an improvement too.

But I would prefer to not add a global variable at all.

When could this check ever evaluate to false?  Compression has
been supported for a long time now.

It appears that a developer would have to modify
SourceLevelDebugPkg\Include\TransferProtocol.h to select an
older protocol revision.

If we really want build configuration of the protocol revision
then we should use a PCD feature flag or a FixedAtBuild PCD to
allow the developer to configure it from the DSC file.

Mike

> -----Original Message-----
> From: edk2-devel [mailto:edk2-devel-bounces@lists.01.org] On
> Behalf Of Ard Biesheuvel
> Sent: Tuesday, August 29, 2017 8:01 AM
> To: Wu, Hao A <hao.a.wu@intel.com>
> Cc: Ni, Ruiyu <ruiyu.ni@intel.com>; edk2-devel@lists.01.org
> Subject: Re: [edk2] [PATCH]
> SourceLevelDebugPkg/DebugAgentCommon: Use global variable for
> revision
> 
> On 29 August 2017 at 14:09, Hao Wu <hao.a.wu@intel.com> wrote:
> > Since some static code checkers complain that comparing two
> macros will
> > generate a constant result, this commit uses a global
> variable to
> > reflect the revision information of the debug agent, rather
> than using a
> > macro directly.
> >
> > Cc: Ruiyu Ni <ruiyu.ni@intel.com>
> > Contributed-under: TianoCore Contribution Agreement 1.1
> > Signed-off-by: Hao Wu <hao.a.wu@intel.com>
> > ---
> >  .../Library/DebugAgent/DebugAgentCommon/DebugAgent.c
> | 7 ++++---
> >  1 file changed, 4 insertions(+), 3 deletions(-)
> >
> > diff --git
> a/SourceLevelDebugPkg/Library/DebugAgent/DebugAgentCommon/Debug
> Agent.c
> b/SourceLevelDebugPkg/Library/DebugAgent/DebugAgentCommon/Debug
> Agent.c
> > index f156fe24db..93af009f96 100644
> > ---
> a/SourceLevelDebugPkg/Library/DebugAgent/DebugAgentCommon/Debug
> Agent.c
> > +++
> b/SourceLevelDebugPkg/Library/DebugAgent/DebugAgentCommon/Debug
> Agent.c
> > @@ -130,7 +130,8 @@ GLOBAL_REMOVE_IF_UNREFERENCED
> EFI_VECTOR_HANDOFF_INFO mVectorHandoffInfoDebugAge
> >    }
> >  };
> >
> > -GLOBAL_REMOVE_IF_UNREFERENCED UINTN mVectorHandoffInfoCount
> = sizeof (mVectorHandoffInfoDebugAgent) / sizeof
> (EFI_VECTOR_HANDOFF_INFO);
> > +GLOBAL_REMOVE_IF_UNREFERENCED UINTN  mVectorHandoffInfoCount
> = sizeof (mVectorHandoffInfoDebugAgent) / sizeof
> (EFI_VECTOR_HANDOFF_INFO);
> > +GLOBAL_REMOVE_IF_UNREFERENCED UINT32 mDebugAgentRevision
> = DEBUG_AGENT_REVISION;
> >
> >  /**
> >    Calculate CRC16 for target data.
> > @@ -1564,7 +1565,7 @@ ReadMemoryAndSendResponsePacket (
> >      // Compression/decompression support was added since
> revision 0.4.
> >      // Revision 0.3 shouldn't compress the packet.
> >      //
> > -    if (DEBUG_AGENT_REVISION >= DEBUG_AGENT_REVISION_04) {
> > +    if (mDebugAgentRevision >= DEBUG_AGENT_REVISION_04) {
> >        //
> >        // Get the compressed data size without modifying the
> packet.
> >        //
> > @@ -2192,7 +2193,7 @@ CommandCommunication (
> >        break;
> >
> >      case DEBUG_COMMAND_GET_REVISION:
> > -      DebugAgentRevision.Revision = DEBUG_AGENT_REVISION;
> > +      DebugAgentRevision.Revision = mDebugAgentRevision;
> >        DebugAgentRevision.Capabilities =
> DEBUG_AGENT_CAPABILITIES;
> >        Status = SendDataResponsePacket ((UINT8 *)
> &DebugAgentRevision, (UINT16) sizeof
> (DEBUG_DATA_RESPONSE_GET_REVISION), DebugHeader);
> >        break;
> > --
> > 2.12.0.windows.1
> >
> 
> Can we *please* fix the tools instead? mDebugAgentRevision has
> external linkage, and so the compiler is forced to perform this
> check
> at runtime instead of at build time. Or at least use a STATIC
> variable
> so the compiler can see that mDebugAgentRevision is never
> modified
> from its default value.
> _______________________________________________
> edk2-devel mailing list
> edk2-devel@lists.01.org
> https://lists.01.org/mailman/listinfo/edk2-devel